Image Retrieval Using Fuzzy $k$-NN on Dominant Rotated Local Binary Patterns
S. Sivakumar, Sarmitha Sathiamoorthy · 2019 International Conference on Smart Systems and Inventive Technology (ICSSIT) · 2019
Day to day progress in technologies, rapidly magnify the size of image archives and thus process of retrieving desired images from huge image archives is one of active research topic and it has wide range of applications. This process is more difficult when the images are affected by illumination conditions, rotations and spatial resolutions. In this paper, local level, rotation and illumination invariant texture information are computed using dominant rotated local binary pattern (DRLB) for texture and histological datasets and fuzzy k-NN with Canberra distance as kernel is applied on DRLBPs for categorizing the images then divergence among database and query images are measured using Canberra distance. The experimental outcomes provide evidence for the efficiency of proposed approach for texture and histological datasets.